A4Vision Revolutionizes Checkpoint Processing, Applying 3D to 2D Facial Identification Database Searching

A4Vision Inc. today announced new 3D facial identification capability to revolutionize search results against existing 2D photo databases. A4Vison adds 3D precision to the 2D data capturing process and employs its ICAO compliant 2D-3D cameras to bring new value to the largest applications of 2D image databases-- border control checkpoints, watch lists, and civil ID programs. Immediate benefits derive from a new robustness to angles that provides more user-friendly capture at borders and other high throughput security points; and the ability to normalize angles to increase result quality in matching subjects to watch list images.

One of the known limitations of current search techniques is an inability to offer robustness to real life angles required when searching large criminal justice watch list databases or verifying a live person against a 2D image on the fly in a high throughput scenario, said Kenneth Moses, AFIS forensic expert and former head of Crime Scene Investigations in San Francisco. A4Visions new 3D capability brings much needed precision and immediate value in time, effort and money saved to 2D database searching or border control verification. Many applications are due to benefit from A4Visions 3D capability applied to 2D facial recognition.

A4Visions new ICAO compliant 2D-3D camera captures 2D and 3D images simultaneously in real-time. After overlaying a 2D photo with the 3D image, then normalizing the captured image to direct frontal image (0 degrees), the normalized image can then be used to verify against passports, or search against 2D image databases. When 3D images captured at checkpoints can be applied at exact angle to 2D database or ID document images, new precision/robustness is achieved in real world scenarios.

A4Visions new 3D capability applied to 2D database searches addresses outstanding problems with facial recognition at checkpoints. When live subjects at checkpoints present facial images that vary from angles captured in 2D photos and stored in 2D databases, both watch list and generic subjects have been difficult to identify; yet both subject types require recognition in order to maintain checkpoint flow. A4Visions new 3D capability applied to 2D facial identification databases establishes a new realm of accuracy. 3D biometric facial identification provides actual, measurable, real-time data about a subjects face, skull and bone structure, enabling higher throughput and improved security on the spot.

Applying A4Visions 3D capability to 2D facial identification searches gains more precision and better matching results, and provides the mechanism for creating 3D image databases.